How do I identify my ideal customer for B2B SaaS?

The Short Answer

Look at the customers you already have who renewed, expanded, and referred someone. Find what those three have in common that your churned customers didn't. That pattern is your ICP. You don't invent it in a workshop, you extract it from the accounts that already worked.

Why Personas Don't Help

Most early SaaS teams build a persona document. Job title, company size, a stock photo, a made-up name, some bullet points about goals and pain points. It feels like progress and it changes nothing, because a persona describes a type of person rather than a buying situation.

"VP of Operations at a 200-person company" doesn't tell you whether they'll buy. "VP of Operations who just absorbed a second team and inherited two systems that don't talk to each other" tells you exactly when to reach them and what to say.

The situation is the ICP. The title is just how you find the situation.

Extract It From Your Own Data

Pull three lists. Customers who renewed and expanded. Customers who churned. Deals you lost. Then look for what separates them.

Check company size, but also check the things nobody puts in a persona. What triggered them to look. Who found you. How long onboarding took. Whether they had someone internally who owned the outcome. How many other tools they were already running.

Usually one variable does most of the work. It might be that everyone who renewed had a dedicated ops person, and everyone who churned expected the tool to replace one. That's not a demographic. That's the actual line between fit and no fit, and you'd never have guessed it in a workshop.

The Trigger Event Matters More Than the Profile

A perfect-fit company that has no reason to change this year is not a prospect. Timing is part of fit.

So write down what happens right before someone buys. A funding round. A new hire in a specific role. A compliance requirement. Hitting a headcount where the manual process breaks. Losing a big customer. Those triggers are findable, and they turn outbound from spray into timing.

Too Broad Is the Default Failure

Almost every early SaaS company defines its ICP too wide, because narrowing feels like shrinking the market. It reads as caution. It's the opposite.

Overwatch had strong technology and a Y Combinator pedigree, and two years after demo day they had no revenue. The market wasn't the problem. Their positioning covered so much ground that no buyer could tell what they were purchasing. Once the offer was cut down to specific high-urgency use cases in cyber and fraud, they landed pilots with Visa, Chase, and NASDAQ and closed a $5M round. Same technology. Narrower target.

Your ICP is a marketing decision, not a policy about who you'll accept money from. You can sell to whoever shows up. But the message that makes anyone show up has to be aimed at one person.

Write It Down in One Sentence

"[Role] at [type of company] who just [trigger] and now needs [outcome]." If your sentence has an "and" joining two different audiences, you haven't finished deciding.

Then test it against your pipeline. If most of your current deals don't match, either the ICP is wrong or your outbound is aimed somewhere else. Both are worth knowing.

Where to Start

Take your last ten customers. Sort them into thriving, fine, and gone. Write one sentence about what was true at signup for each group. The difference between column one and column three is your ICP, and it took you twenty minutes instead of a quarter.

If you're pre-revenue and don't have customers to learn from yet, you're defining the ICP rather than extracting it. That is what the CoBuilder builds with you: the buyer profile, the trigger events, and the offer aimed at them. Seven days free with full access. And if you're a technical founder who built the product first and is now trying to figure out who it's for, that's the exact situation we wrote this page for.